simhash.py 文件源码

def fit(self, X, y=None):
        """
        Parameters
        ----------
        X : {array, sparse matrix}, shape (n_samples, n_features)
            List of n_features-dimensional data points. Each row
            corresponds to a single data point.
        Returns
        -------
        self : object
            Returns self.
        """
        from simhash import compute
        self._fit_X = X = check_array(X, accept_sparse='csr')

        n_features = X.shape[1]

        def _scale_hash_32_64bit(indices):
            return indices*((2**64-1)//2**32-1)

        hash_func = self.hash_func

        hashing_table = np.array(
                [hash_func(el, 0) for el in range(n_features)],
                dtype='uint64')

        shash = []
        for idx in range(X.shape[0]):
            # get hashes of indices
            mhash = hashing_table[X[idx].indices]
            if self.hash_func_nbytes == 32:
                mhash = _scale_hash_32_64bit(mhash)
            shash.append(compute(mhash))
        _fit_shash = np.asarray(shash, dtype='uint64')
        self._fit_shash = _fit_shash
        self._fit_shash_dict = {val: key
                                for key, val in enumerate(self._fit_shash)}